Indirect Carbonation of Victorian Brown Coal Fly Ash for CO2

Sep 9, 2014 - as carbonate from Victorian brown coal fly ash has been examined. Victorian .... Brown coal is the single largest energy source in the s...
0 downloads 0 Views 5MB Size
Download PDF
Article pubs.acs.org/EF

Indirect Carbonation of Victorian Brown Coal Fly Ash for CO2 Sequestration: Multiple-Cycle Leaching-Carbonation and Magnesium Leaching Kinetic Modeling Tahereh Hosseini,† Cordelia Selomulya,† Nawshad Haque,‡ and Lian Zhang*,† †

Department of Chemical Engineering, Monash University, Clayton, GPO Box 36, Victoria 3800, Australia Process Science and Engineering Division, CSIRO, Clayton, Victoria 3168, Australia

‡

ABSTRACT: In this paper, a closed-loop multistep process which allows leaching and precipitation of magnesium and calcium as carbonate from Victorian brown coal fly ash has been examined. Victorian brown coal fly ash has a distinctively high concentration of alkaline earth metals and low amounts of aluminum and silica. The main objective here is to clarify the dissolution kinetics of magnesium and calcium in regenerative ammonium chloride and subsequent carbonation of dissolved cations. Instead of a once-through test with fresh ammonium chloride, multiple locked circuits were adopted to assess the leaching capability of regenerated ammonium salt, as well as the accumulation of impurities upon the recycling and reuse of the leaching agent. As has been revealed, upon increasing cycles of ammonium chloride use, the extraction yields of both target cations decreased significantly. Their extraction by ammonium chloride was favored by the presence of the oxide form in the original ash sample, with the extraction of calcium occurring much faster than that of magnesium. Both phenomena were in agreement with the thermodynamic equilibrium prediction on the lowest Gibbs function for the dissolution of oxides, especially calcium oxide in ammonium chloride solution. Carbonation results dropped gradually upon the increase in the cycle number; meanwhile, the size and morphology of precipitates were changed from the first to last cycle. By fitting the observed results with a shrinking core model, it was shown that the extraction of Mg2+ followed a pseudo-second-order reaction with a nonconstant ammonium chloride concentration in the film layer on the surface of a solid particle. The activation energy of 20.7 kJ mol−1 was obtained for the dissolution of magnesium from both Hazelwood fly ash and pure MgO in ammonium chloride solution.

1. INTRODUCTION Mineral carbonation or mineralization is one of the most promising processes for carbon capture, storage, and utilization (CCSU) to combat the climate change. In this process, carbon dioxide (CO2) reacts with alkaline oxides or hydroxides, specifically magnesium and/or calcium in natural minerals or industrial waste, to convert into thermodynamically stable carbonates, thus avoiding the necessity of the costly monitoring of CO2 leakage during transportation and storage.1 Mineral carbonation technologies can be divided into singlestep (direct) and multiple-step (indirect) categories. The single-step method involves a direct reaction of feedstock material with CO2 in a single reactor maintained at controlled temperature and pressure which generally fall in the ranges 100−500 °C and 10−20 bar, respectively.2 In contrast, the multistep indirect carbonation is initiated by the dissolution of a mineral species in an aqueous medium to extract the alkaline earth metals within it. The resulting cations are subsequently carbonated through bubbling the CO2-containing flue gas into the leachate. Both leaching and carbonation take place under relatively mild conditions, e.g., 20−80 °C and 1−5 bar. The leaching media for the dissolution of calcium and magnesium in minerals are either strong or weak acidic reagents, ammonium salts, or alkaline solutions.3 Of those, the ammonium extraction process appears to be a favorable technique, given its potential for reagent recovery and the relatively large selectivity of magnesium and calcium over other elements,4 as demonstrated by the reactions below: © 2014 American Chemical Society

Extraction: Mg/CaO + 2NH4Cl → Mg/CaCl 2 + H 2O + 2NH3(g) (1)

Carbonation: Mg/CaCl 2 + CO2 + 2NH3(g) + H 2O → Mg/CaCO3 + 2NH4Cl

(2)

To date, various types of minerals rich in alkaline earth metals, either natural species or man-made industry waste, have been tested for their carbonation propensities. These include serpentine,5 olivine and wollastonite,6 fly ash derived from coal combustion and municipal solid waste incineration (MSWI),3,7 blast furnace, and steelmaking slag.1,8 Brown coal fly ash is a potentially appropriate source, as it is rich in magnesium and/ or calcium. The fly ash is generated on-site with CO2 together in a power plant. It possesses a particle size distribution with the majority falling in the micron or even submicron scale, thereby requiring no communition prior to utilization. In addition, the conversion of fly ash from zero value waste into value-added high-purity carbonates could create extra income stream for power plants, thereby offsetting the carbon tax to be implemented in the carbon-constrained future.3,9−12 The Received: June 26, 2014 Revised: September 9, 2014 Published: September 9, 2014 6481

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

Table 1. Waste Minerals Elemental Quantification (XRF) composition (%) Hazelwood fly ash Yallourn fly ash

SiO2

Al2O3

Fe2O3

CaO

MgO

Na2O

K2O

P2O5

SO3

5.82 8.92

3.01 5.92

14 42.3

32.4 9.4

29.3 27.9

0.2 0.56

0.17 0.16

0.41 0.11

12.8 2.84

capacity to sequester CO2 for fly ash depends directly on the proportion of binary oxides (CaO and MgO) and/or hydroxides (Ca(OH)2 and Mg(OH)2) contained in the waste matrix. For example, a fly ash containing a high percentage of free magnesium in oxide form can be carbonated more readily than a fly ash with the same magnesium content but in the form of silicates. Thus, it is difficult to directly compare the efficiency of various fly ashes for mineral carbon sequestration.13 Brown coal is the single largest energy source in the state of Victoria, Australia, meeting >85% of its electricity need. The combustion of brown coal yields up to 1.3 million tons of fly ash annually, nearly all of which was simply dumped in ash ponds. Victorian brown coal fly ash has a chemical composition dominated by magnesium, calcium, iron, and sulfur and is classified as being a strongly alkaline fly ash that is not suitable as cement’s additive.3 Due to its high moisture content, the brown coal combustion generates a high CO2 emission rate relative to the high-rank black coal and natural gas in Australia.14 A cost-effective carbon capture process is pivotal for the sustainability of brown coal in the carbon-constrained future. Moreover, with the continuous increase in the amount of fly ash generated, it incurs the increasing demand on the use of vast land for landfills, which also contaminates the soil, ground, and water simultaneously.15 On the assumption of a full conversion of magnesium and calcium to carbonate for a typical Victorian brown coal fly ash containing around 32% CaO and 29% MgO (Table 1), the carbonation reaction is expected to capture 278 kg of CO2 per ton of raw fly ash, which is equivalent to a total amount of 0.36 million tons of CO2 captured per annum. Sun et al. (2012) compared the CO2 sequestration capacity of Victorian brown coal fly ash with other fly ashes and found 264 kg of CO2 per ton of raw fly ash at 60 °C, 10 bar, and 1 h reaction time, which is remarkably higher than lignite and black coal fly ashes.3 Although this number is much smaller than the total amount of CO2 released, it is expected that the industrial mineral wastes can act as a supplement to the natural minerals for CO2 capture. Considering the broad variation of fly ash properties, the methods developed in the literature are generally not applicable to different samples. To date, the majority of studies on fly ash utilization have focused on direct carbonation under a high CO2 partial pressure with a rather long reaction time.16−18 However, reports on utilization of coal fly ash are limited with

the research target set only to optimize the conditions for oncethrough leaching and carbonation stages. The integration of leaching and carbonation stages and the recovery of leaching reagent are yet to be discussed. Moreover, the mechanisms underpinning the dissolution of magnesium upon leaching of fly ash are rarely reported.19 Few papers reported leaching kinetics of different elements including Al, Fe, and Ca and trace elements like Cr, Zn, and As from coal fly ash.20,21 Elemental leaching from fly ash is a complex process, which involves dissolution, diffusion, adsorption, and mineral precipitation. Leaching of ash takes place through dissolution of constituents inside or on the surface of ash particles and transport through the pore structure to the surrounding solution. For solid−fluid reactions, the shrinking core kinetic model22 has been used widely. Ranjitham et al. (1990),23 Raschman (2000), 24 and Atashi et al. (2010) 25 have successfully applied this model to the leaching kinetics of calcined magnesite in ammonium chloride with a concentration of about 1 M. Paul et al. (2004) indicated that the kinetics of acid consumption for different types of Turkish fly ashes consist of an initially fast process followed by a slower period.19 To reiterate, all of these studies only focused on the once-through fresh leaching reagent. In this study, two types of Victorian brown coal fly ash were tested in multiple cycles of the leaching-carbonation closed loop using regenerative ammonium chloride as the leaching reagent. The aim is to clarify the extraction and carbonation mechanisms of magnesium (as well as calcium) in an industrially relevant process. That is, rather than the use of fresh reagent for once-through investigation, a multicycle experiment has been conducted to reveal the recyclability of a regenerative reagent, ammonia chloride. As a comparison to fly ash, pure oxide compounds of predominant elements (MgO, CaO, and Fe2O3) and their mixtures were also examined in the multicycle mode to evaluate the influence of impurities in fly ash on the extraction of target oxides. Furthermore, a detailed modeling on the leaching of magnesium from pure oxide and fly ash was conducted by assessing the applicability of various models, so as to provide an accurate model for future scale-up.

2. MATERIALS AND EXPERIMENTAL METHODS 2.1. Materials Preparation. Two coal fly ash samples were collected as dry powders from the electrostatic precipitator in International Power Hazelwood and Energy Australia Yallourn power plant located at the Latrobe Valley, Victoria, Australia. Once being delivered to the laboratory, each fly ash sample was initially washed with water with a liquid to solid (L/S) mass ratio of 10 to remove the unburnt carbon and water-soluble species such as sodium and potassium sulfate. Subsequently, the water-washed fly ash samples were dried in an oven at 120 °C overnight and crushed mildly below 150 μm prior to use. Pure MgO and CaO samples were prepared by calcination of magnesium carbonate and calcium carbonate (purchased from Sigma-Aldrich), respectively. The temperature and time for calcination in a muffle furnace were fixed at 800 °C and 12 h, respectively.23 The resulting MgO and CaO obtained were analyzed by thermogravimetric analysis (TGA) to ensure a complete calcination of the carbonates. Pure Fe2O3 was purchased from Sigma-Aldrich in

Table 2. NH4Cl Concentration at the Start Point of Each Cycle for MgO, MgO + CaO, MgO+Fe2O3, and Hazelwood Fly Ash NH4Cl concentration (M) cycle

MgO

MgO + CaO

MgO+Fe2O3

Hazelwood Fly ash

1 2 3 4 5

4 3.47 3.02 2.81 2.67

4 3.52 3.04 2.84 2.73

4 3.58 3.01 2.86 2.76

4 3.38 2.99 2.85 2.72 6482

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

Figure 1. XRD spectra for Hazelwood fly ash and Yallourn fly ash. 2.4. Single Leaching Conditions. For each run, 10 g of a sample in dried powder form was mixed with 60 mL, 4 M ammonium chloride at a L/S ratio of 6. The liquid to solid ratio and ammonium chloride concentration were fixed throughout this study. Instead, a broad range of temperature (25, 40, 60, and 80 °C) and time (10−60 min) was varied and three replications were carried out for each condition. The temperature was not increased further due to the limitation of operation at higher temperatures in industry and to avoid evaporation of aqueous solutions that might increase the concentration of ammonium chloride, thus causing the dissolved ammonium chloride to precipitate back in leaching residue. In addition, the study on leaching behavior of calcined magnesite carried out by Ranjitham and Khangaonkar (1990) showed an insignificant change of leaching progress at temperatures above 80 °C.23 Batch leaching tests were performed in a glass sealed beaker equipped with two connections, one for feeding air of 1 L·min−1 and another tube for releasing ammonia vapor that is then trapped in a conical flask containing distilled water. The reaction temperature (40− 80 °C) was controlled by immersing the reactor in a thermostatcontrolled water bath. A magnetic stirrer bar with a stirring speed of 350 rpm enabled the solution to be fully agitated with minimal spillage. The resulting residue after filtration was dried at 120 °C overnight in an oven, weighed, and quantified by XRF for elemental compositions to determine the leaching percentages of individual elements, particularly magnesium and calcium. The resulting ammonium water was titrated by acetic acid (1 M) to determine the amount of the ammonia recovered from the leaching step. 2.5. Single Carbonation Conditions. Following the leaching experiment, the resulting leachate was subsequently carbonated under the conditions of the pH of the 9−108 through the doping of ammonia−water (NH3, generated from the leaching step), a continuous injection of pure (grade 4.5) CO2 at 15 L·min−1 for 20 min. The amount of ammonia added in the carbonation tank is exactly the same as that evaporated from the leaching step to ensure a closure of the whole process in NH3 balance. These conditions were optimized in our previous works. According to reactions 1 and 2, the Mg2+ and Ca2+ cations in the leachate are expected to fully precipitate out as solid carbonate, which in turn results in the regeneration of ammonia chloride to be used in the next round. The resulting carbonate was filtered and dried overnight in the oven, and its mass was recorded.

oxide form. In addition, the analytical grade ammonium chloride was purchased from Merck with a purity of 98%. For the comparison to fly ash samples, the pure oxide and their mixtures were also tested on multiple leaching-carbonation cycles, including pure MgO, mixtures of MgO + CaO (mass ratio 1:1), and mixtures of MgO + Fe2O3 (mass ratio 1:4). These ratios were selected on the basis of molecular weights and weight percentages of these metals in the two fly ash samples. 2.2. Materials Characterization. The elemental composition of a raw sample, its leaching residues, and precipitated carbonates were determined by a precalibrated X-ray fluorescence spectroscopy (XRF, Spectro iQ II). About 3−4 g of a representative sample was finely ground and stored in a sample holder for the XRF analysis. The leaching percentage of each element was calculated on the basis of the difference between its mass in raw washed sample and solid residue after drying. The produced carbonates were dried overnight at 105 °C and weighted to determine the conversion. The mineralogical composition of raw fly ash and few leaching residues and carbonates was determined by X-ray diffraction analysis (XRD, Rigaku, Miniflex), under a scanning speed of 1° min−1 from 2 to 90°, 40 kV and 15 mA. The peak identification was achieved by search-match function in the JADE software. Scanning electron microscopy (SEM) was employed for morphology observation of carbonate precipitates. The sample powder was dispersed into a carbon-taped sample holder stab and platinum coated. Each carbonate precipitate was characterized by randomly selecting 3− 4 fields of view and examining all the fly ash particles observed within the selected fields. The SEM microscope used is a JEOL JSM-7001F equipped with energy dispersive X-ray spectroscopy (EDX). SEM imaging studies were performed at 15 kV at a working distance of 10 mm. 2.3. Thermodynamic Equilibrium Calculation. The different species and phases, which were known to exist, were specified into the reaction equation module of HSC Chemistry 7.1 to calculate multicomponent equilibrium compositions for the heterogeneous systems. The heat capacity, enthalpy, entropy, and Gibbs energy of a single species and reaction systems of pure substances were provided by the built-in database in the HSC Chemistry.26 The Gibbs free energy of the reactions of various Mg- and Ca-bearing species with ammonium chloride were calculated in a wide temperature range. 6483

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

2.6. Five-Cycle Leaching-Carbonation Conditions. Once a cycle of leaching-carbonation is finished, the regenerated ammonia chloride is tested again for a total of five cycles in this study. The experiments for all the leaching and carbonate exactly follow the above-mentioned conditions, except that a small amount of fresh hydrochloric acid (2 M) was doped into the regenerated chloride to reduce its pH back to the original value of 4.5. The increase in the pH of ammonia chloride is attributed to the accumulation of unreacted alkali and alkaline earth metal cations after carbonation.

equilibrium fraction of magnesium leached in 1 h from around 20% at 25 °C to 32% at 80 °C. The increasing trend of temperature is consistent with the observation of Ranjitham and Khangaonkar (1990) on the leaching of calcined magnesite with temperature up to 80 °C. However, they achieved around 43% magnesium extraction at 80 °C and 30 min reaction time with an L/S of 20. The lower pH of their solution emerging from using a significant amount of leaching agent is the likely cause of discrepancy.23 At a given temperature, the leaching fraction of magnesium increased exponentially over time, reaching its maximum at around 30 min. The leaching of magnesium was nearly ceased from 30 min onward. The probable explanation for this phenomenon is the crystallization of ammonium chloride and the possible reaction of free Mg2+ with OH− from ammonium hydroxide on the surface of ash at high pH, creating a passive layer on the surface of particles to block the continued leaching of magnesium. To validate the above proposed hypothesis, the thermal decomposition of leaching residue obtained from conditions of 80 °C and 1 h for the use of pure MgO was conducted by TGA, at a heating rate of 5 °C/min from room temperature to 800 °C. Figure 3 illustrates a three-step mass loss for the residue tested. The first loss was commenced before 200 °C, which can be assigned to hydrate; the second one at around 330 °C is attributed to the decomposition of ammonium chloride crystals which is 338 °C for the pure compound;27 and the last one at about 490 °C was due to the decomposition of magnesium hydroxide. A significant decrease in weight of leaching residue after washing with water was another proof for the existence of ammonium chloride crystal, which accounted for around 48% of the total residue. Such a phenomenon has been confirmed by Wang et al. (2012)28 who has observed that the crystallization of NH4Cl occurs from the NH4Cl-rich solution when the concentration of MgCl2 within it is increased up to 2.5−2.7 mol·L−1. The concentration of MgCl2 in the leachate achieved in our study reaches 2.62 mol·L−1, which falls in the above range. The theoretical and practical ammonia recovery as a function of leaching time and temperature are presented in Figure 4. The theoretical ammonia recovery was calculated according to eq 1, based on the amount of magnesium cation extracted. Increasing temperature favors the recovery of ammonia. This was expected, since the solubility of ammonia will decrease at elevated temperatures. However, the practical ammonia recovery is much lower than the corresponding theoretical value, indicating that a certain fraction of evaporated ammonia gas is still present in the leachate. This may not be a big issue, as the ammonia is essential for the subsequent carbonation of the dissolved Ca2+ and Mg2+ cations. 3.3. Five Leaching-Carbonation Cycles. 3.3.1. Leaching Results. The leaching percentages of magnesium in pure oxide compounds upon the reuse of ammonia chloride in the five cycles are presented in panel a of Figure 5, and panel b shows the results of the two fly ashes tested. For pure oxide mixtures, the extraction yield of magnesium is relatively constant for the three mixtures, regardless of the cycle number. This is an indicator of the independence of magnesium extraction on the mass of MgO in a solid sample mixture. For all of the pure oxide mixtures, the total mass of each solid remained the same while the mass of MgO is different because its content is different for the three mixtures. For each pure species, upon the increase on cycle number, the extraction yield of magnesium

3. RESULTS AND DISCUSSION 3.1. Properties of Fly Ashes. Table 1 tabulates XRF results in oxide form for as-received and washed Hazelwood and Yallourn fly ash samples. The major elements are Mg, Ca, and Fe. Hazelwood fly ash is rich in MgO and CaO (29.3 and 32.4%, respectively) and also includes 14% Fe2O3, whereas the predominant elements in Yallourn fly ash are Fe2O3 and MgO (42.3 and 27.9%, respectively) with only 9.4% CaO. Figure 1 depicts the XRD spectra for Hazelwood and Yallourn fly ash. As can be seen, periclase (MgO) is the only Mg-bearing crystal species in Hazelwood fly ash, which is accompanied by Cabearing species including anhydrite (CaSO4), lime (CaO), calcium ferrite (CaFe2O4), and silicate. Only two major crystalline phases were found in Yallourn fly ash, with magnesia ferrite showing the strongest intensity and quartz with a medium peak height. Interestingly, calcium (9.4% in its oxide form) in Yallourn fly ash was present as an amorphous structure that was undetectable by XRD analysis. 3.2. Single Leaching Results. The solid lines in Figure 2 illustrate the leaching percentage of pure MgO as a function of

Figure 2. Effect of time and temperature on leaching of magnesium. Solid lines represent leaching from pure MgO, and dashed lines refer to Hazelwood Fly ash.

temperature and time. In contrast, the dashed lines show the magnesium leaching percentage from Hazelwood fly ash. The error bars refer to standard deviations from measured values of three replicates. The nearly identical results for magnesium extraction from both pure MgO and Hazelwood fly ash can be confirmed for the same condition. This substantiates the presence of the majority of magnesium as free oxide in Hazelwood fly ash and the insignificant influence of impurities and the other forms of magnesium in this fly ash, if any. The leaching temperature was influential, as it improved the 6484

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

Figure 3. Derivative mass loss of dried residue against temperature running TGA (T = 80 °C, t = 1 h).

increasing the cycle number. The calcium leaching yields from two fly ash samples reach 32% for the first cycle and only 10% at the fifth cycle for Hazelwood fly ash, and from 37 to 3% for Yallourn fly ash during the three cycles tested. This was attributed to different modes of occurrence of calcium in two fly ash samples, i.e., an abundance of a mixture of anhydrite, lime, and calcium ferrite in Hazelwood fly ash, and undetectable amorphous calcium alumino-silicate in Yallourn coal fly ash. The dissolution rates of both magnesium- and calciumbearing species in ammonium chloride were compared using thermodynamic equilibrium calculation (Figure 7). For the dissolution of MgO in ammonium chloride, its ΔG turns negative from 80 °C onward, suggesting the spontaneity of the forward reaction to occur at this temperature. This is consistent with our results for the leaching of pure MgO and Hazelwood fly ash summarized in Figure 2. The lower magnesium leaching percentage of Yallourn fly ash rich in magnesia ferrite can also be explained by the positive ΔG for this reaction from room temperature to 200 °C in Figure 7. For both calcium oxide and portlandite (Ca(OH)2), their dissolution ΔG is always lower than the other Ca-/Mg-bearing compounds, substantiating the spontaneous dissolution of these two species in ammonium chloride at every temperature. This supports the experimental observation for higher calcium extraction percentage from the mixture of MgO + CaO than the two real fly ash samples. The decreasing trend upon the five cycles suggests that the ammonium chloride leaching capability has declined upon recycling. It is hypothesized that increasing the concentration of dissolved magnesium and calcium ions in leachate favored two side reactions: 1 - reaction of ammonium chloride with dissolved Mg2+ cation to form a complex precipitate, 2 ammonium chloride crystallization during carbonation. The second hypothesis has been proven by the leaching residue mass loss profile upon heating in Figure 3. To further support the first hypothesis, XRD analysis results for the leaching residue and carbonate from the first cycle of multiple-cycle leaching-carbonation of pure MgO are presented in Figure 8. The presence of a Mg−Cl complex in the untreated leaching

Figure 4. Theoretical and practical ammonia recovery. Dashed lines are related to theoretical ammonia recovery, and solid lines correspond to practical ammonia recovery.

drops slowly from ∼35% in the first cycle to 30% in the fifth cycle. A similar observation was confirmed for Hazelwood fly ash. However, for Yallourn fly ash, the leaching yield of magnesium reached only 25% in the first three cycles. Its leaching and carbonation were fully stopped from the third cycle (results for the fourth and fifth cycles are not shown). This should be due mainly to a strong association of most of magnesium with iron in the form of magnesia ferrite in this fly ash. Accordingly, the ammonia salt was too weak to break the strong association of MgO and Fe2O3. The leaching percentage of calcium from the MgO + CaO mixtures has also been compared to the two fly ash samples. The results are presented in Figure 6. Compared to a rather constant extraction yield of around 80% for calcium from the MgO + CaO mixture, the leaching yield of calcium out of two fly ashes is much lower and decreases dramatically upon 6485

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

Figure 5. Leaching % for Mg during five-cycle leaching-carbonation at optimum conditions (T = 80 °C, t = 30 min). (a) Pure MgO, MgO + CaO, and MgO+Fe2O3. (b) Hazelwood and Yallourn fly ash.

residue was verified. XRD results show that the peak characteristic of Mg−Cl complex disappeared after water washing. Similarly, the ammonium chloride crystals in carbonate products were confirmed too, which again vanished upon water washing. Similar results were confirmed for the leaching residues and carbonates generated in the other cycles (XRD data not shown). Moreover, on the basis of the assumption of producing 1 mol of Mg−Cl complex upon consumption of 1 mol of ammonium chloride, the weight loss after washing can be further derived for the fractionation of ammonium chloride complex coupled with dissolved magnesium. The results are depicted in Figure 9. For the pure MgO and its mixture with CaO, the crystallization extent of ammonium chloride is more significant than the fraction of solid complex formed. The crystallization extent also increases noticeably upon increasing the cycle number. Such a trend was however not found for the solid complex. In contrast, the solid complex formed has a comparable and even higher fraction than that of ammonium chloride crystal for the pure mixture of

Figure 6. Leaching % for Ca during five-cycle leaching-carbonation at optimum conditions (T = 80 °C, t = 30 min), in MgO + CaO and Hazelwood and Yallourn fly ash.

Figure 7. Thermodynamic equilibrium calculation for different Mg- and Ca-bearing species in ammonium chloride solution. 6486

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

Figure 8. XRD results for first-cycle leaching residue and carbonate before and after washing for pure MgO.

Figure 9. Loss percentage of NH4Cl at each cycle breaking down to crystallized and complex formed for MgO, MgO + CaO, MgO + Fe2O3, and Hazelwood fly ash. Water washing was conducted at room temperature, L/S = 10 and t = 1 h.

MgO + Fe2O3 and Hazelwood fly ash. This suggests that Fe2O3 and other impurities in fly ash are in favor of the combination of ammonium and MgO into complex solid. This explains the

extremely low extraction yield for magnesium out of MgO + Fe2O3 mixture and Yallourn coal fly ash. Moreover, the concentration of ammonium chloride at each cycle were also 6487

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

Figure 10. Carbonation % for Mg during five-cycle leaching-carbonation at optimum conditions (T = RT, t = 20 min). (a) Pure MgO, MgO + CaO, and MgO + Fe2O3. (b) Hazelwood and Yallourn fly ash.

of MgO + CaO, Ca2+ cation showed a slightly low carbonation extent, with a similar trend of decreasing slightly from 90% in the first cycle to 76% in the last cycle. The Ca2+ cation in the leachate of Hazelwood fly ash shows the same calcium carbonation behavior as the mixture of oxide (MgO + CaO) in the first and second cycles. However, its carbonation extent was decreased sharply after the second cycle and reached 32% at the fifth cycle. Similarly, the Yallourn fly ash carbonation percentage is the lowest among the samples tested, dropping from 40% to a negligible value at the third cycle. The availability of calcium and magnesium cations in a solution is one of the significant factors influencing their carbonation.29 The competition between ions present in solution can influence the nature and morphology of precipitated carbonates.30 Figure 12 and Table 3 illustrate a suite of typical SEM images and the respective elemental analysis for single spots in the carbonate precipitates from the first and last cycles of (a) pure MgO, (b) MgO + CaO, (c) MgO + Fe 2 O 3 , and (d) Hazelwood fly ash. Being complementary to the SEM observation, XRD results on the major Mg- and Ca-bearing carbonate minerals are summarized in Table 4. In the case of pure MgO demonstrated in Figure 12a, its first cycle carbonate is dominated by discrete rosettes of magnesite (MgCO3). Under the natural (i.e., ambient) conditions, hydromagnesite or other metastable phases such as nesquehonite and lansfordite can be formed, depending on the availability of Mg2+ ions in solution in relation to the availability of other cations such as Ca2+.31 In addition, the transformation of nesquehonite to magnesite or amorphous magnesium carbonate at 70−100 °C was reported by Hollingbery et al. (2010), which could occur during drying in the oven at a temperature of 110 °C.32 Instead, the last cycle precipitates showed aggregates of small round particles. Such a discrepancy should be caused by the decreased concentration of magnesium cation in the leachate in the last cycle, which preferred to precipitate into small crystals that are too dilute to agglomerate together. This is consistent with the observation of Case et al. (2011) that concluded a higher concentration of magnesium ion in solution along with adding substrate increases the nucleation rate of magnesium precipitates.33 Upon the initial central nuclei or hydroxide formed, the formation of magnesite and its nucleation growth can be accelerated significantly.34 Regarding the case of the MgO + CaO mixture, its carbonate precipitates in Figure 12b suggest a polymorph of round

calculated and summarized in Table 2. As can be seen that, for all of the four leachates examined, ammonium chloride concentration was decreased gradually from first cycle to last cycle. It is another direct evidence for the loss of ammonium chloride from the solution upon recycling. 3.3.2. Carbonation Results. Figure 10 demonstrates the carbonation percentage of Mg from the leachates of (a) different mixtures of pure oxides and (b) two fly ashes. Panel a indicates that, irrespective of the oxide mixture type, the carbonation percentage of magnesium reaches nearly 100% at the first cycle, and drops gradually upon the increase in the cycle number. For the two fly ash samples in panel b, magnesium in the Hazelwood Fly ash leachate was nearly fully carbonated in the first cycle. Its carbonation yield, however, was dropped quickly to only 50% in the fifth round. This is clearly faster than the pure oxide compounds shown in panel a. Moreover, the performance of Yallourn fly ash leachate is even worse, with the carbonation degrees for Mg2+ cations decreasing sharply from 31% in the first cycle to nearly zero in the third cycle. These results link with the aforementioned phenomenon on the loss of ammonium and magnesium into solid complex and crystallization species in the leaching residues. The carbonation percentages of calcium in MgO + CaO, Hazelwood fly ash, and Yallourn fly ash are presented in Figure 11. Compared to Mg2+ cation in the leachate from the mixture

Figure 11. Carbonation % for Ca during five-cycle leachingcarbonation at optimum conditions (T = RT, t = 20 min) in MgO + CaO and Hazelwood and Yallourn fly ash. 6488

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

Figure 12. SEM images of carbonate obtained from the first and last cycles of multiple-cycle leaching-carbonation: (a) Pure MgO; (b) MgO + CaO; (c) MgO + Fe2O3; (d) Hazelwood fly ash.

The crystalline structures of carbonate precipitates from the first and last cycles of MgO + Fe2O3 are present in Figure 12c. In comparison to the precipitates from MgO leachate (Figure 12a), the presence of iron in the leachate apparently changed the morphology of the magnesium carbonate particles to a relatively porous structure. However, there is not a significant difference observed between the first and last cycles. The smallest drop in the carbonation extent of MgO + Fe2O3 from the first to last cycle in comparison to other mixtures is consistent with a negligible change in carbonation morphology of magnesium trough recycling. This can be attributed to an inferior concentration of magnesium ion in the leachate and a low tendency of iron to compete with magnesium to form carbonate or precipitate in the pH range 7−9.5. Carbonate precipitates obtained from first and last cycle leaching-carbonation of Hazelwood fly ash are present in panel d of Figure 12. As can be seen, significant changes occur for the particle shape as the cycle number shifts from first to last. First cycle precipitates show randomly oriented polycrystalline aggregates, while at the last cycle formation of submicron rod shape particles coalesced and settled in the irregular shaped agglomerates was observed. XRD results for the first cycle precipitates showed incorporation of both magnesium and calcium in the crystal lattice in the form of calcite magnesian ((Ca,Mg)CO3). The calcium sulfate peak in the last cycle XRD spectra shows an affinity of the calcium ion to react with a high concentration of sulfate ion in the leachate to from solid precipitate. Bischoff and Fyfe (1968) reported inhibition of calcite formation strongly by the presence of magnesium and less strongly by sulfate ions.35 The increased concentration of

Table 3. SEM-EDX Elemental Analysis of Representative Precipitate Particles Obtained from the First and Last Cycles of Five-Cycle Leaching-Carbonation of Pure MgO, MgO + CaO, MgO + Fe2O3, and Hazelwood Fly Ash elemental composition (wt %) sample

cycle

pure MgO MgO + CaO MgO + Fe2O3 Hazelwood fly ash

first last first last first last first last

Ca

34 14

16 23

Mg 26 4 2 20 24 30 20 14

Fe

S

Cl

O

2

21 8 8 13 20 5 13 7

53 68 56 53 56 65 51 54

particles for the first cycle while in the last cycle an aggregate of rhombohedral platy particles was observed. EDX mapping indicates the scattering of magnesium spots among the calciumrich particles. Along with this, strong peaks were formed for magnesite and lansfordite in the XRD spectra. All of these suggest the incursion of magnesium ion into the internal layers of the calcite lattice. The sensitivity of calcium carbonate crystallization to the presence of magnesium ions has been reported by several investigators.30,35 Roques and Girou (1974) have reported that calcium carbonate precipitation was markedly reduced by the presence of a small amount of magnesium ions in the solution. The ions of Mg stabilize amorphous, unstable, and hydrated phases and thus decrease the quantity of well crystallized carbonates.36

Table 4. Different Mg- and Ca-Bearing Minerals Detected by XRD for the First and Last Cycles of Carbonation Obtained from Multiple-Cycle Leaching-Carbonation of Pure MgO, MgO + CaO, MgO + Fe2O3, and Hazelwood Fly Asha cycle first cycle

last cycle

pure MgO

MgO + CaO

magnesite (MgCO3)

magnesite (MgCO3)

hydromagnesite [Mg5(CO3)4(OH)2·4(H2O)]

Lansfordite [MgCO3·5(H2O)] calcite (CaCO3) magnesite (MgCO3)

MgO + Fe2O3

magnesian calcite [(Ca, Mg)CO3]

magnesite (MgCO3)

magnesite (MgCO3)

Lansfordite [MgCO3·5(H2O)] calcite (CaCO3) a

Hazelwood fly ash

magnesite (MgCO3)

calcite (CaCO3) calcium sulfate (CaSO4)

Oven drying at 110 °C may cause the formation of magnesite from the conversion of nesquehonite. 6489

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

slowest step, the following expression of the shrinking core model can be used to describe the dissolution kinetics:

interfering ions concentration in leachate upon recycling clearly affected the morphology of carbonate. 3.4. Kinetic Modeling of Magnesium Leaching. In comparison to an obvious time dependence of the leaching of magnesium in Figure 2, the leaching of calcium in the form of oxide is much faster, which was proven in Figure 13 to be

1 − (1 − XB)1/3 =

(4)

Vice versa, in the case that the diffusion of magnesium ion through the ash layer is the rate controlling step, the following equation can be used: 1−

2DMBVA 2 XB − (1 − XB)2/3 = t = K dt 3 ρB αr0 2

(5)

In eqs 4 and 5, XB is the fraction of solid reacted, KC is the kinetic constant, MB is the molecular weight of the solid, CA is the concentration of dissolved lixiviant A in the bulk of solution, ρB is the density of solid, a is the stoichiometric coefficient of the reagent in the leaching reaction, r0 is the initial radius of the solid particle, t is the reaction time, D is the diffusion coefficient in the porous product layer, and Kr, Kd are the rate constants which are calculated from eqs 4 and 5, respectively.37 The rate of dissolution of magnesium was tested against diffusion control (eq 4) and chemical control (eq 5). For this purpose, the left sides of these equations were plotted with respect to reaction time. The fitting level of these models was evaluated using correlation coefficient (R2) values. The slopes of these plots were used as the apparent rate constants (Kr and Kd). Table 5 summarizes the diffusion control and chemical

Figure 13. Effect of time and temperature on leaching of calcium from Hazelwood fly ash.

Table 5. Dissolution Rate and R2 Using the Reaction Controlled Model, Diffusion Control Model, and Mixed Reaction and Diffusion Control Model

finished in less than 5 min, irrespective of the leaching temperature. Clearly, the leaching of calcium in the form of oxide is mainly thermodynamically controlled, whereas its reaction rate is fast enough. In other words, for the coexistence of magnesium and calcium in a sample such as Hazelwood fly ash, the leaching of magnesium out of the solid matrix is the limiting step for the overall extraction. In this sense, the kinetic modeling for magnesium was further performed in this paper. For the noncatalytic reaction of particles with surrounding fluid, two simple idealized models can be considered, the progressive-conversion model (PCM) and the shrinking unreacted-core model (SCM). In the PSM, it is assumed that the reactant enters and reacts through the particle at all times; thus, solid reactant is converted continuously and progressively throughout the particles. Instead, the SCM assumes the reaction occurs first at the outer skin of the particle and the reaction zone then moves into the solid, leaving behind partly or completely converted material and inert solid which can precipitate back on the surface of the particle. Thus, at any time during the reaction, there exists an unreacted core of material which shrinks in size. Back to Figure 2, it is clear that the leaching process was controlled either by diffusion of reactant through the solution boundary, through a solid product layer, or by the surface chemical reaction rate.22 Therefore, the shrinking core model for magnesium leaching is considered here. In the model, a solid particle B immersed in a fluid A reacts with the fluid by eq 3: aA(Fluid) + B(Solid) → Products

K CMBCA = k rt ρB αr0

reaction controlled model

diffusion controlled model

1 − (1 − XB)1/3

1 − (2/3)XB − (1 − XB)2/3

T (°C)

R2

Kr × 103 (min−1)

R2

Kd × 104 (min−1)

25 40 60 80

0.7482 0.7543 0.7054 0.5779

1.142 1.333 1.663 1.728

0.8458 08551 0.7973 0.6104

2.944 4.010 5.984 6.676

control models of leaching reaction in terms of rate constant and correlation coefficient. Clearly, the correlation coefficients are mostly below 0.8, which are too poor to be accepted. In other words, the above approaches based on the classical shrinking core model are inaccurate. A careful examination of the leaching results in Figure 2 reveals that the magnesium concentration in the bulk liquid solution initially increased very fast and then leveled off at an equilibrium value in each temperature. Furthermore, the previous discussions have confirmed the crystallization of ammonium chloride and the formation of complex mixture as solid residues. They grow steadily upon the gradual dissolution of magnesium into the leachate. As a result, the resistance against diffusion of ions through the boundary should increase gradually, which eventually leads to the full termination of the reaction. For simplification, the leaching of magnesium was simulated by two steps, the first 30 min for step 1 and the remaining 30 min for step 2 where the leaching is rather stopped. The kinetic data for the first step were matched well by the fitting of an empirical pseudo-second-order reaction with nonconstant ammonium chloride concentration in the outer surface of MgO particles, as presented in eq 6. This model is

(3)

We assume MgO as solid B and ammonium chloride as fluid A, and according to eq 1, the stoichiometric coefficient of a can be set as 2. When the first-order surface chemical reaction is the 6490

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

plots of ln(K) versus the reciprocal of temperature for the pseudo-second-order reaction model (eq 6) further indicate an activation energy of 20.7 kJ mol−1 for the leaching of magnesium in the first 30 min, as demonstrated in Figure 15.

also consistent with the stoichiometric coefficient of ammonium chloride in eq 1 and the decreasing trend of the magnesium leaching results from multiple-cycle leachingcarbonation of pure MgO and Hazelwood fly ash presented in Figure 5. To reiterate, the reason for a decreased extraction of magnesium upon multiple cycles is due to the decreased concentration of free ammonium chloride. To reflect this point, the concentration of ammonium chloride was introduced into the left-hand side of eq 6, as a nonconstant term and as a function of time and temperature. By further dividing both the left- and right-hand sides by CA, the concentration of ammonium chloride at time t, the effect of the initial concentration of ammonium chloride on leaching is further added into the model. CA can be substituted in terms of initial concentration and conversion. K M (1 − XB)−1 = C B = K rt ρB αr0 CA 0 − 2CB0XB

(6)

The diffusion control model presented in eq 5 was proven to fit the kinetic data for the second step satisfactorily. In summary, Figure 14 shows the R2 values of two proposed models for (a) the first step and (b) the diffusion control model for the second step. One can see the much better correlation coefficients for the newly developed model. The Arrhenius

Figure 15. Arrhenius plot for leaching of MgO with ammonium chloride using the pseudo-second-order reaction model.

Ranjitham et al. (1990),23 Raschman (2000),24 and Atashi et al. (2010)25 reported activation energies of 43.2, 48.5, and 42.2 kJ mol−1, respectively, for the leaching of calcined magnesite, all of which are relatively higher than our result. This may be due to a much smaller fly ash size examined here. Luo et al. (2013)38 concluded activation energies of 32, 28, and 19 kJ mol−1 for dissolution of aluminum, calcium, and iron, respectively, from calcined Chinese coal fly ash in hydrochloric acid, which is clearly comparable with present work. Figure 16 further

Figure 16. K values fitting the second-order model for different cycles of Hazelwood fly ash.

demonstrates a linear dependence of the K value on the initial/free concentration of ammonium chloride for the leaching of magnesium from Hazelwood fly ash/pure MgO. This further confirmed the applicability of this model for predicting magnesium leaching kinetics in the multiple-cycle leaching-carbonation.

Figure 14. Plots for (a) pseudo-second-order reaction for the first 30 min of reaction and (b) product layer diffusion for the last 30 min of reaction. 6491

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

(6) Gerdemann, S. J.; O’Connor, W. K.; Dahlin, D. C.; Penner, L. R.; Rush, H. Ex Situ Aqueous Mineral Carbonation. Environ. Sci. Technol. 2007, 41, 2587−2593. (7) Hong, K. J.; Tokunaga, S.; Kajiuchi, T. Extraction of heavy metals from MSW incinerator fly ash using saponins. J. Hazard. Mater. 2000, 75, 57−73. (8) Sun, Y.; Yao, M. S.; Zhang, J. P.; Yang, G. Indirect CO2 mineral sequestration by steelmaking slag with NH4Cl as leaching solution. Chem. Eng. J. 2011, 173 (2), 437−445. (9) Nyambura, M. G.; Mugera, W. G.; Felicia, P. L.; Gathura, N. P. Carbonation of brine impacted fractionated coal fly ash: implications for CO2 sequestration. J. Environ. Manage. 2011, 92, 655−664. (10) Montes-Hernandez, G.; Pérez-López, R.; Renard, F.; Nieto, J. M.; Charlet, L. Mineral sequestration of CO2 by aqueous carbonation of coal combustion fly-ash. J. Hazard. Mater. 2009, 161 (2−3), 1347− 1354. (11) Huijgen, W. J. J.; Witkamp, G. J.; Comans, R. N. J. Mineral CO2 sequestration by steel slag carbonation. Environ. Sci. Technol. 2005, 39 (24), 9676−9682. (12) Bonenfant, D.; Kharoune, L.; Sauvé, S.; Hausler, R.; Niquette, P.; Mimeault, M. CO2 Sequestration potential of steel slags at ambient temperature and pressure. Ind. Eng. Chem. Res. 2008, 47 (20), 7610− 7616. (13) Bobicki, E. R.; Liu, Q.; Xu, Z.; Zeng, H. Carbon capture and storage using alkaline industrial wastes. Prog. Energy Combust. Sci. 2012, 308, 302−320. (14) Johnson, T. R. In Future Options for Brown Coal based Electricity Generation − the role of IDGCC, Proceedings of the ANZSES Destination Renewable Conference, Melbourne, November 26th−29th, 2003; 371−380. (15) Wu, C. Y.; Yu, H. F. Extraction of aluminum by pressure acidleaching method from coal fly ash. Trans. Nonferrous Met. Soc. China 2012, 22, 2282−2288. (16) Huang, K.; Inoue, K.; Harada, H. Leaching of heavy metals by citric acid from fly ash generated in municipal waste incineration plants. J. Mater. Cycles Waste Manage. 2011, 13, 118−126. (17) Kersch, C.; Pereto; Ortiza, S.; Woerlee, G. F.; Witkampa, G. J. Leachability of metals from fly ash: leaching tests before and after extraction with supercritical CO2 and extractants. Hydrometallurgy 2004, 72, 119−127. (18) Soco, E.; Kalembkiewicz, J. Investigations of sequential leaching behaviour of Cu and Zn from coal fly ash and their mobility in environmental conditions. J. Hazard. Mater. 2007, 145, 482−487. (19) Paul, M.; Seferinoglu, M.; Ayçik, G. A.; Sandström, A.; Paul, J. Acid leaching of coal and coal-ash: kinetics and dominant ions. American Chemical Society, 228th National Meeting Conference proceeding, 2004. (20) Seidel, A.; Zimmels, Y. Mechanism and kinetics of aluminum and iron leaching from coal fly ash by sulfuric acid. Chem. Eng. Sci. 1998, 53 (22), 3835−3852. (21) Zhu, Z. Characterization and modeling of toxic fly ash constituents in the environment. Ph.D. Thesis, University of Tennessee, 2011. (22) Levenspiel, O. Chemical Reaction Engineering; John Wiley and Sons: New York, 1972. (23) Ranjitham, A. M.; Khangaonkar, P. R. Leaching Behaviour of Calcined Magnesite with Ammonium Chloride Solutions. Hydrometallurgy 1990, 23, 177−189. (24) Raschman, P. Leaching of calcined magnesite using ammonium chloride at constant pH. Hydrometallurgy 2000, 56, 109−123. (25) Atashi, H.; Fazlollahi, F.; Tehranirad, S. Leaching Kinetics of Calcined Magnesite in Ammonium Chloride Solutions. Aust. J. Basic Appl. Sci. 2010, 4, 5956−5962. (26) Pickles, C. A. Thermodynamic modelling of the formation of zinc- manganese ferrite spinel in electric arc furnace dust. J. Hazard. Mater. 2010, 179, 309−317. (27) http://www.chem.unep.ch/irptc/sids/OECDSIDS/12125029. pdf.

4. CONCLUSION A comprehensive investigation on the leaching propensity of dominant oxides (MgO and CaO) in two different types of brown coal fly ash (Hazelwood and Yallourn) using ammonium chloride has been conducted. Apart from the parametric investigation over the influence of time and temperature on the once-through leaching experiment, five closed leachingcarbonation loops have been performed to explore the reusability of the leaching reagent, as well as the accumulation of impurities upon recycling. The optimum conditions for leaching of magnesium from pure MgO within the range studied here were found to be at 80 °C, a reaction time of 30 min, a liquid/solid ratio of 6, and an ammonium chloride concentration of 4 M. Identical results for magnesium leaching yields have been observed for all samples, except for Yallourn coal fly ash with a chemically stable magnesia ferrite (MgFe2O4). The leaching and carbonation percentage of magnesium and calcium decreased upon increasing cycle number, due to the loss of ammonium chloride by crystallization and its interaction with dissolved magnesium forming complex precipitates. The leaching of calcium in the form of oxide is much faster than magnesium, which can be finished in less than 5 min at a given temperature. Instead, the leaching of magnesium in Hazelwood fly ash was slow in the first 30 min, following a pseudo-secondorder reaction with a nonconstant ammonium chloride concentration, while the kinetic data for the second step showed a good fit to the diffusion-controlled process. The activation energy for the leaching of MgO was found to be about 20.7 kJ mol−1 from the start of reaction toward 30 min after the reaction, which is consistent with the value obtained for the reaction control kinetic modeling.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +61-3-9905-2592. Fax: +61-3-9905-5686. E-mail: lian. [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This project is supported by the Faculty of Engineering of Monash University for a 2012−2013 seed grant. The first author is also grateful to Monash Research Graduate School (MRGS) for a Ph.D. scholarship.

■

REFERENCES

(1) Eloneva, S.; Teir, S.; Salminen, J.; Fogelholm, C. J.; Zevenhoven, R. Fixation of CO2 by carbonating calcium derived from blast furnace slag. Energy 2008, 33, 1461−1467. (2) Sanna, A.; Dri, M.; Hall, M. R.; Maroto-Valer, M. Waste materials for carbon capture and storage by mineralisation (CCSM) − A UK perspective. Appl. Energy 2012, 99, 545−554. (3) Sun, Y.; Parikh, V.; Zhang, L. Sequestration of carbon dioxide by indirect mineralization using Victorian brown coal fly ash. J. Hazard. Mater. 2012, 209−210, 458−466. (4) Teir, S.; Revitzer, H.; Eloneva, S.; Fogelholm, C. J.; Zevenhoven, R. Dissolution of natural serpentinite in mineral and organic acids. Int. J. Miner. Process. 2007, 83, 36−46. (5) Maroto-Valer, M. M.; Kuchta, M. E.; Zhang, Y.; Andrésen, J. M.; Fauth, D. J. Comparison of physical and chemical activation of serpentine for enhanced CO2 sequestration. Prepr. Pap. - Am. Chem. Soc., Div. Fuel Chem. 2004, 49 (1), 373−375. 6492

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493

Energy & Fuels

Article

(28) Wang, D.; Li, Z. Study of Crystallization Kinetics of Ammonium Carnallite and Ammonium Chloride in the NH4Cl-MgCl2-H2O System. Ind. Eng. Chem. Res. 2012, 51, 2397−2406. (29) Han, Y. S.; Hadiko, G.; Fuji, M.; Takahashi, M. Effect of flow rate and CO2 content on the phase and morphology of CaCO3 prepared by bubbling method. J. Cryst. Growth 2005, 276, 541. (30) Reddy, M. M.; Nancollas, G. H. The crystallization of calcium carbonate: iv. The effect of magnesium, strontium and sulfate ions. J. Cryst. Growth 1976, 35, 33−38. (31) Frost, R. L.; Hales, M. C.; Locke, A. J.; Kristof, J.; Horvath, E. Vagvolgyi, Controlled rate thermal analysis of hydromagnesite. J. Therm. Anal. Calorim. 2008, 92, 893−897. (32) Hollingbery, L. A.; Hull, T. R. The Thermal Decomposition of Huntite and Hydromagnesite - A Review. Thermochim. Acta 2010, 509, 1−11. (33) Case, D. H.; Wang, F.; Giammar, D. E. Precipitation of Magnesium Carbonates as a Function of Temperature, Solution Composition, and Presence of a Silicate Mineral Substrate. Environ. Eng. Sci. 2011, 28, 881−889. (34) Langmuir, D. Stability of Carbonates in the System MgO-CO2H2O. J. Geol. 1965, 73, 73. (35) Bischoff, J. L.; Fyfe, W. S. Catalysis, inhibition and the calcitearagonite problem. Am. J. Sci. 1968, 266, 65−79. (36) Roques, H.; Girou, A. Kinetics of the formation conditions of carbonate tartars. Water Res. 1974, 8, 907−920. (37) Dehghan, R.; Noaparast, M.; Kolahdoozan, M. Leaching and kinetic modelling of low-grade calcareous sphalerite in acidic ferric chloride solution. Hydrometallurgy 2009, 96, 275−282. (38) Luo, Q.; Chen, G.; Sun, Y.; Ye, Y.; Qiao, X.; Yu, J. Dissolution kinetics of aluminum, calcium and iron from circulating fluidized bed combustion fly ash with hydrochloric acid. Ind. Eng. Chem. Res. 2013, 52, 18184−18191.

6493

dx.doi.org/10.1021/ef5014314 | Energy Fuels 2014, 28, 6481−6493